Grant Number: 5R01CA106744-04
PI Name: MURPHY, SUZANNE P.
Project Title: Evaluating Errors in Reported Use of Dietary Supplements
Abstract: DESCRIPTION (provided by applicant): Dietary supplements are widely used (up to 70% of some groups report regular use), often frequently and sometimes at high doses. As a result, vitamin and mineral supplements make a substantial contribution to nutrient intakes for many Americans, and intakes of any type of supplement (including botanicals) may have important effects on chronic disease outcomes. However, methods to characterize supplement use at either the individual or group level have seldom been evaluated. Although errors in the measurement of food intake have been extensively studied, little is known about the errors associated with self-reports of supplement use. The proposed study will evaluate the two most common methods of collecting dietary supplement use data: (1) A self-administered short questionnaire that asks about use of a specific list of supplements over the past year; and (2) An open-ended interviewer-administered recall of recent supplement use (past two weeks or past month). Both methods will be compared to a reference inventory method, which measures intake based on a quarterly inventory of supplement containers, collected for one year. The open-ended recalls will also be compared to daily dairies, a reference measure for shorter-term supplement use. Data collection will span one year for each of 500 subjects, representing five ethnicities and both genders. Three-fifths of the sample (300) will provide supplement use data for the inventory, diaries, and both recall methods. Two-fifths of the sample (200) will serve as control subjects and will only complete the short questionnaire at the beginning and end of the study. Data analyses will compare results between methods and between time periods, and quantify errors in types of supplement used, frequency of use, and doses. Daily, weekly, and quarterly (seasonal) within person variation in supplement use will be quantified. The effect of using default values for brand, dose, and frequency will also be investigated. The results will (1) inform the choice of instruments to be used for national surveys and large cohort studies; (2) suggest approaches to minimize errors in self-reports of supplement use; and (3) permit the development of appropriate adjustments that can be used during analyses to account for variation and biases in the self-reported data. Back to Grants Page